| The development characteristics and susceptibility evaluation of landslide and slide geological hazards are very important to guide the construction of regional human engineering.Based on the data collection,field investigation and experimental analysis,the susceptibility of landslide and slide geological disasters in Fugu County was evaluated by combining ArcGIS,R language and SPSS software,and some conclusions were obtained:(1)Among the landslide and slide geological disasters in Fugu County,there were 92 collapses,and most of them were small and rock collapses,accounting for 65.22%and 79.35%of the total collapse disasters,respectively.There were 28 landslides,and the development types were mainly small landslides,soil landslides and shallow landslides,accounting for 71.43%,89,29%and 78.57%of the total landslide damage points,respectively,and there was a pattern of "gradually increasing from west to east" in the distribution.The main controlling factors for the development of landslide and slide geological hazards in Fugu county are five categories:landform,type of rock and soil mass,groundwater action,rainfall and human engineering activities.(2)Ten evaluation factors of landslide and slip geological hazards in Fugu County were determined,including elevation,slope aspect,slope,curvature,rainfall,distance from road,distance from water system,normalized vegetation index,type of rock and soil mass,topographic humidity index.(3)Established the frequency ratio,random forests(FR-RF),frequency ratio,logistic regression(FR-LR),frequency ratio,logistic regression,random forests(FR-LR-RF)three coupling model,with slope unit and grid unit separately fugu county collapsed-slippery geological hazards assessment,and calculate the three models in two kinds of evaluation unit collapse-slide of geological disasters in the proportion of high-high rock.The results show that under the slope unit,the number of landslid'e and slip geological hazards in the extremely high prone areas accounts for 77.24%,67.58%and 88.52%,respectively.Under the grid unit,the number of landslide and slip geological disasters in the high-extremely high prone areas accounted for 71.66%,63.33%and 78.33%,respectively.(4)The subject characteristic curve method and Kappa coefficient method were used to compare the models of the two evaluation units.The results showed that the prediction rate curve AUC value of FR-LR model,FR-RF model and FR-LR-RF model of the slope unit was 0.012,0.016 and 0.017 higher than that of the three models of the grid unit,respectively.According to the calculation of Kappa coefficient,the coefficients of FR-LR model,FR-RF model and FR-LR-RF model under the slope element are 0.083,0.017 and 0.025 higher than those under the raster element,respectively.In the two evaluation elements,the accuracy of the three coupling models is higher than that of the two coupling models,and the accuracy of the FR-LR-RF model under the slope element is 0.02 higher than that of the raster element. |
PDF Full Text Request